How people's attention flows on the web
The web is a social artifact.
Here is a selected collection of related items:
While we’re a good-natured group, we have serious respect for designers and inventors alike, and we want to make sure the design community – including everyone involved with Quirky – has all of the information needed to understand what’s going on.
Quirky is a business that sources ideas for products from inventors via the internet and, having used its ‘community’ to select designs, puts these ideas into a process potentially leading to production.
Quirky is accusing OXO of appropriating a design from one of its inventors and ‘took to the streets’ to claim ‘justice for Bill Ward’. This could have been a public relations mess for OXO.
Have a read of how they handled it. Cool and calm. They headed off Quirky’s little-person versus big-person marketing stunt by schooling Quirky on how intellectual property works.
Let me know what you think on Twitter
While writing my last post I came across this apposite article on the Content Marketing Institute’s website. The chart is from a report conducted by a business called ExactTarget. The sample used is rather opaque: ‘US online population aged 18 or older’ and the question itself is somewhat unclear but it is the interpretation of the data that is most amusing.
The above indicates the disparity between what marketers think the public wants versus what the public says they want:
“When customers were asked where their favorite brands should invest their marketing time and resources to improve customer loyalty, just 6 percent answered that they want related (or helpful) content from a brand. In fact, more than double that number of consumers actually wanted to receive product-related content.”
Hmm. Where does that leave this description of the point of content marketing from elsewhere on their site:
“Basically, content marketing is the art of communicating with your customers and prospects without selling. It is non-interruption marketing. Instead of pitching your products or services, you are delivering information that makes your buyer more intelligent. The essence of this content strategy is the belief that if we, as businesses, deliver consistent, ongoing valuable information to buyers, they ultimately reward us with their business and loyalty.”
Hence all this hype from marketers around this notion of ‘brands as publishers’ on the web. But the chart above is telling us in a slightly confused way that:
- People who have given their permission by a business to be emailed are happy to receive email from that business (no surprise!)
- The company website should host appropriate content when they want more information…
- … and this content should be about products…
- … supported by in-store help
My interpretation of this is that the ‘consumers’ who responded were thinking about finding answers to product-related questions.
The writer of the aforementioned article comes to a rather different conclusion quoting a chap from the company that produced the report:
“If you really take a look at the research, consumers are actually desperate for content from brands, but like marketers they are fixated on channels,” says Rohrs. “When consumers ask for more email and Facebook, they are asking for helpful content through those channels. What are marketers going to put into those channels… air?”
Um. Are they really ‘desperate’? Is that what these people labelled as ‘consumers’ are asking for? It sounds as though people want to know about new products and when they are aware of them want to be able to find out the details easily and in their own time, hence the lack of desire for marketing messages through other social platforms.
The response to marketing through Facebook is not necessarily to do with ‘helpful content’ but is likely an indication of a willingness to consider ‘offers and promotions’. Have a read of this survey where the top responses to the question “I connect with brands on Facebook and other social networks” were:
- For the games, contests and promotions
- To learn about new products
- To show I’m a fan
Note that none of these are asking for ‘helpful content’ that is non-product-related.
Reminds me of this quote from The Ad Contrarian:
“We don’t get them to try our product by convincing them to love our brand. We get them to love our brand by convincing them to try our product.”
The conclusion of the original article is most ironic:
“Make sure you take your own personal behavior out of the equation before you make any type of content marketing channel decisions.”
Yes, such as being convinced that the service you are selling to clients is one that ‘consumers’ actually want!
Let me know what you think on Twitter
Following on from my last post in which I emphasised the importance of evidence to support your assertions I have provided the following for review.
I’ve looked at the analytics of all sorts of websites. The chart below shows a typical distribution of page views to posts on a site set-up as a ‘content destination’:
This particular site followed a blog-style format on which several posts a week appeared. The salient points are:
- In over 600 posts in over two-and-a-half years only one post came close to breaking 10,000 page views
- The second most popular post received just under half as many page views
- 60% of the posts received less than a hundred visits over the same time-scale
- New visitors made up greater than 80% of the audience, which means there were no real readers
- People visited once, frequently looking at just one page, and did not come back
I do not have the data to tell how many were ‘readers’ as this was not defined at the time but it was clear the format was not working at attracting people. This was a shame as waht was being posted was great! There were no pushy sales messages. It was that elusive thing: ‘quality content’ professionally produced. No one cared.
It was this observation that set me off investigating this pattern, especially as clients were being advised to spend money on creating and publishing content that was not specifically about their core products or services. This project was the first of several, for very different brands, that demonstrated the same pattern of audience behaviour, for which I have the data. This formed the basis of my view that generalising about ‘brands as publishers’ on the web, along with the volume of content production that implies, is not a sustainable approach. Yet many marketers are recommending precisely this even though it will not help meet business objectives. This distribution is not a one-off and is a pattern followed by all content destinations, the difference is that successful sites receive massively higher amounts of attention and possess a regular readership, something that requires both significant resources and a perception of independence by the audience. At higher volumes the pattern still exists, but the critical mass of readers sustains even the long-tail of what’s being created.
This approach is useful in certain circumstances. Friends of mine run a very successful music label and as genre experts the things they post about their artists are not only popular with their customers, they’re directly attributable to sales. This does not mean their posts do not follow a similar distribution to the above. The difference is they write about their products for people who are likely to buy them so the cost of producing the posts is covered and they are very careful about ensuring this is the case; their time is precious, as should be yours.
What frustrates me is how often basic and observable behaviours are ignored and how often assumptions lead to the repetition of the same mistakes over and over again. Partly this is to do with vested interests; there are a lot of agencies out there who carry out content marketing functions who see this an opportunity to move from tactical and valuable one-off activity to an expensive and retained model of constant generation of new ‘content’. Yet these people remain wilfully ignorant of how attention flows on the web, despite the data being easy to collect and interpret.
Here are some examples of brand ‘content destinations’ that are completely ineffective at reaching an audience. The mistake in each case is to assume ‘if we build it they will come’ and that the ‘brand’ is strong enough to own a website with a readership in line with the definition of a ‘reader’ given in my last post.
Tech Page One from Dell:
“Technology. Business. Lifestyle. Those are the three pillars of Tech Page One’s content streams. Under those categories, we’ve commissioned and curated blog posts, in-depth articles, and infographics to keep you current on the major conversations in tech.”
The major question that should have been asked here is why, given the limited time people have, would they turn to a Dell-owned property for tech opinion when there is Engadget, Gizmodo, The Verge, et al. Their Twitter profile is following more people than they have followers. The design is poor as is the use of stock photography. It is doomed.
mb! from Mercedes:
“mb! by Mercedes-Benz is a magazine that documents and reflects contemporary culture and a clever, fun way of living. If you like personal and intelligent stories, reports and interviews as well as inspiring visuals reflecting the current Zeitgeist, then you’ll love mb!”
mb! from Mercedes by contrast is nicely designed and focused on a particular lifestyle with which the brand presumably wishes to be associated. In some ways this makes it worse as it is likely an even greater input of time has gone into something no-one is reading. Once again, there are any number of lifestyle destinations that do the same thing more authentically, with a more frequently updated and diverse array of feature articles. The danger is in the association; I don’t actually genuinely believe Mercedes Benz as a business has any interest in skateboarding or any of the other topics featured on this website.
What is bonkers is that Mercedes Benz seems to also be paying for a site with video content on a similar theme called the Avant Garde Diaries.
Frankly, I would love to run these sites for Mercedes Benz, but I could not in good conscience take their money because I know what little return is likely being generated. When someone at Mercedes Benz cottons on these sites are doomed.
Take a look at the number of video views generated by Avant Guard Diaries YouTube channel. The top rated video has about 4,500 views, the second 2,000, then it drops off a cliff (remember the chart above!). There are countless videos here with less than a hundred views!
mb! is likely receiving no more than a couple of thousand visits a month in the UK(*). How many are readers?
Energy Forecaster from uSwitchforBusiness:
“That’s why Business Juice started Energy Forecaster. We want businesses to be well-informed, well-prepared and most importantly, involved, when it comes to energy. The site is full of content that we hope will help businesses like yours get to grips with the importance of energy, and on-going changes to prices, legislation, government schemes and more.”
This site is a resource for businesses. The notion of ‘guides’ to selected topics is a good one. The idea of regular blogging by a panel of ‘industry experts’ is a good one. The idea of hosting this all on your own content destination is diabolically appalling and dooms this site to failure.
The worst part of this for me is that I found this site via a blog post (since removed though slides available here) about agile content strategy in which it is clear a lot of effort has gone into creating it. There was a lot of stuff around ‘thinking about the user’ with no consideration that this user is an entirely imaginary person and that, because of how attention works on the web, it is going to take vast resources to develop a readership for this content.
Apparently the site launched in May. As of writing the affiliated LinkedIn Group has 29 members, the Facebook Page has 11 Likes and the Twitter profile has only slightly more followers than followed. Many of the profiles look less than human. Doomed.
These situations have all occurred because of the vested interests that arise around a brand having a presence on the web without properly understanding whether anyone is going to be genuinely interested in coming back to these websites time and again.
What is depressing is that by collaborating with independent sites that already are destinations a far better effect could be achieved, far more simply, with far less overhead and allow for tactical campaigns that actually support measurable objectives.
tl;dr - if you are reading a post about ‘content marketing’ and there is no data to support the examples given then what you are reading is an anecdote. This is useless from a decision-making point-of-view. Challenge people on this, do not just accept their assertions.
(*) This figures were based on the anonymised data of the websites UK households are actually visiting taken from the records of UK ISPs.
Let me know what you think on Twitter
According to Percolate founder Noah Brier, in the next few years brands should expect to post 40 to 60 pieces of meaningful — not just promotional — content a day.
Why, I wonder? To what end? Who has the time to read this stuff? Truth is almost none of these ‘meaningful’ items will ever be seen by human eyes; at least not without a massive and ongoing investment of resources with little in the way of a directly measurable return.
I take issue with the notion of brands as publishers if publishing is not the core business function. There seems to be an assumption that a high volume of publishing to a brand owned and controlled destination is desired by web users. There has been more excitability of late with SEOs piling into this area as Google has put the frighteners on other techniques for gaming the SERPs; a regular stream of items is assumed beneficial for the purposes of ‘link building’.
The theory sounds logical. People want to be informed about the things they are interested in. Search engines reward ‘fresh’ and popular pages. By creating your own pages of this type on a domain you own, based on research, a ‘brand’ can attract people with these interests.
The Content Marketing Institute says it is the:
“… technique of creating and distributing relevant and valuable content to attract, acquire, and engage a clearly defined and understood target audience – with the objective of driving profitable customer action.”
Focus on that last point and keep it in mind as you read this. The great thing about the web is that if you have a theory about what should be done you can collect a lot of data to prove your point. What is significant is how few authors of articles that plug these services actually bother to do this and thus support their position with anything independently verifiable. For instance if you claim ‘guest blogging is an opportunity to build authority and credibility’ then you should be able to show a hundred examples of you doing this exact thing and a thousand more examples of other people doing the same, all with a measurable outcome.
What I want to test is whether this is actually viable at the scale being promoted in the quote that opens this post.
The assumptions I’m making are as follows:
- A brand publishes items to web pages to fulfill business objectives, be it to affect perception, better inform a prospect of the features of a product, to deal with common customer questions, etc. This is the ‘marketing’ part of ‘content marketing’
- To achieve this objective a significant number of people must read these pages, or the messages carried by these pages must be discussed by people to whom your potential customers will listen
- These pages we’re discussing are not product descriptions at the point-of-sale, they are additional pieces of marketing such as that which appears on the ‘company blog’, a micro-site, sub-domain or similar
- Given this costs money to both produce and then support there is a point at which the number of people that see it are too few in number for it to be cost-effective to bother devoting the resources to producing it
- Attention, like pretty much every other behaviour on the web, follows a long tail distribution; for a page to be looked at it has to have links pointing to it
- A ‘page view’ is apropos of nothing and no indication that a page has been properly looked at; readers form a tiny minority of visitors
I suggest reading Identifying Your Real Readers by Thomas Baekdal, who explains this in great detail. He is well worth subscribing to as his posts are excellent.
A reader is someone who actually reads the article, which means:
- They need to scroll from the top to the bottom in a time-frame indicative of someone actually paying attention to what’s on the page
- They need to read several pages in a session or across multiple sessions on different devices, not just visit a single page never to return
- They need to be a repeat visitor; this tells you they are finding what’s on offer of significant value to want to come back to refer to what they have seen and to read more
- Finally they need to have not been referred from somewhere else, they should be your readers, not someone else’s readers who are popping by for a quick look; referrals might turn into readers, but they are not yet a reader on that first visit to a single page
A destination wesbite is about readers not visitors or page views. It’s a really difficult challenge to develop and sustain such a space. Not even professional publishers are succeeding at making it work profitably.
In the next part of this post I will cover some specific examples to see if this can work for a ‘brand’ rather than a ‘publisher’.
Let me know what you think on Twitter
What are some attributes of these networks that will help us make predictions? Is it number of followers? Is it engagement of followers? Is it what time you tweet? Is it who else is tweeting at the same time?
Another look at how news spreads on Twitter. What I question about these studies that claim to help understand when to post is that nowhere does it mention the effect of the content of the post on its propensity to ‘travel’.
That is surely the ineffable question to which everything else is peripheral. But then that is not something that can be predicted by an algorithm.
Let me know what you think on Twitter